How we model revenue at risk (without pretending it is exact)

We combine category benchmarks, traffic proxies, and issue severity to produce a directional range, not a promise.

When crawl data is thin, we narrow claims and widen uncertainty so the report stays honest.

Use the estimate to rank projects by upside, not to litigate decimal places with finance.